Cognitive rehabilitation for people with schizophrenia and related conditions (Review)

Transcription

1 Cognitive rehabilitation for people with schizophrenia and related conditions (Review) McGrath J, Hayes RL This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2009, Issue 1

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY BACKGROUND OBJECTIVES METHODS RESULTS DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES Analysis 1.1. Comparison 1 COGNITIVE REHABILITATION versus PLACEBO COGNITIVE PROCEDURE, Outcome 1 Mental state: Average BPRS (endpoint score, high=poor) Analysis 1.2. Comparison 1 COGNITIVE REHABILITATION versus PLACEBO COGNITIVE PROCEDURE, Outcome 2 Leaving the study early Analysis 1.3. Comparison 1 COGNITIVE REHABILITATION versus PLACEBO COGNITIVE PROCEDURE, Outcome 3 Specific cognitive domains: 1a. Attention Analysis 1.5. Comparison 1 COGNITIVE REHABILITATION versus PLACEBO COGNITIVE PROCEDURE, Outcome 5 Specific cognitive domains: 2a. Memory - logical memory (high=good) Analysis 1.6. Comparison 1 COGNITIVE REHABILITATION versus PLACEBO COGNITIVE PROCEDURE, Outcome 6 Specific cognitive domains: 2b. Memory - visual reproduction (high=good) Analysis 1.7. Comparison 1 COGNITIVE REHABILITATION versus PLACEBO COGNITIVE PROCEDURE, Outcome 7 Specific cognitive domains: 3a. WAIS subtests - object assembly Analysis 2.1. Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 1 Mental state: 1. Average BPRS (endpoint score, high=poor) Analysis 2.4. Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 4 Leaving the study early Analysis 2.5. Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 5 Self esteem: Average Rosenberg score (low=poor) Analysis 2.6. Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 6 Specific cognitive domains: 1a. Cognitive flexibility Analysis 2.8. Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 8 Specific cognitive domains: 2a. Memory tests Analysis 2.9. Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 9 Specific cognitive domains: 2b. Memory tests - dual span Analysis Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 10 Specific cognitive domains: 3a. Planning - Six Elements Analysis Comparison 2 COGNITIVE REHABILITATION versus OCCUPATIONAL THERAPY, Outcome 12 Specific cognitive domains: 4. WAIS subtest - digit span WHAT S NEW HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST SOURCES OF SUPPORT NOTES INDEX TERMS i

3 [Intervention Review] Cognitive rehabilitation for people with schizophrenia and related conditions John McGrath 2, Robyn L Hayes 1 1 School of Occupational Therapy, La Trobe University, Bundoora, Australia. 2 Queensland Centre for Schizophrenia Research, The Park Centre for Mental Health, Wacol, Australia Contact address: Robyn L Hayes, School of Occupational Therapy, La Trobe University, Bundoora, Victoria, 3083, Australia. r.l.hayes@latrobe.edu.au. Editorial group: Cochrane Schizophrenia Group. Publication status and date: Edited (no change to conclusions), published in Issue 1, Review content assessed as up-to-date: 8 March Citation: McGrath J, Hayes RL. Cognitive rehabilitation for people with schizophrenia and related conditions. Cochrane Database of Systematic Reviews 2000, Issue 3. Art. No.: CD DOI: / CD Background A B S T R A C T Researchers have consistently found that people with schizophrenia score more poorly than others on a wide array of cognitive tasks and that these deficits persist even when the illness is in remission. The perceived impact of cognitive impairment on day-to-day functioning has led to the development of cognitive rehabilitation techniques intended to remedy these impairments, and thus improve the functioning of people with schizophrenia. Objectives To determine whether the use of cognitive rehabilitation techniques is associated with improvement in people with schizophrenia and related conditions. Search methods Electronic searches of Biological Abstracts ( ), the Cochrane Schizophrenia Group s Register of trials (2000), EMBASE ( ), LILACS ( ), MEDLINE ( ), PsycLIT ( ) and SCISEARCH (1997) were undertaken. References of all identified studies were handsearched and the first author of each included trial was contacted. Selection criteria The reviewers included randomised trials investigating the impact of cognitive rehabilitation on people with schizophrenia, or related conditions, compared to a placebo intervention, another intervention or standard treatment. This version of the review does not include comparisons of different types of cognitive rehabilitation. Data collection and analysis Relevant studies were identified and data extracted reliably by two reviewers working independently. In studies which did not specify the reasons for participants dropping out, we assumed that those who had dropped out had experienced no change in the outcome measures. Where possible, Peto odds ratios (OR) and mean differences (MD) were calculated with their 95% confidence intervals (CI). 1

4 Main results Three small studies met the inclusion criteria. Two compared cognitive rehabilitation to a placebo intervention (total n=84), and the other to occupational therapy (n=33). Although cognitive rehabilitation was as acceptable as placebo and occupational therapy, with low attrition in both groups, no effects were demonstrated on measures of mental state, social behaviour, or cognitive functioning. An effect, in favour of cognitive rehabilitation, on a measure of self-esteem (Rosenberg Self-Esteem Scale, MD 6.3 CI ) is worthy of replication in any future trials. Authors conclusions Data are inconclusive and provide no evidence for or against cognitive rehabilitation as a treatment for schizophrenia. P L A I N L A N G U A G E S U M M A R Y Cognitive rehabilitation for people with schizophrenia and related conditions Synopsis pending. B A C K G R O U N D Researchers have consistently found that people with schizophrenia score more poorly than others on a wide range of cognitive tasks and that these deficits persist even when the illness is in remission (Gold 1993). The problems that people with schizophrenia experience in day-to-day functioning, and the failure of existing psychosocial rehabilitation approaches (e.g., social skills training) to improve participants day-to-day functioning have been attributed, at least in part, to cognitive impairments (Bellack 1992). The perceived impact of cognitive impairment has led to the development and evaluation of cognitive rehabilitation techniques intended to remedy these impairments, and thus improve the functioning of people with schizophrenia. Cognitive rehabilitation (CR) involves the retraining of basic-level processes in various inter-related cognitive domains such as memory, attention, speed of processing and abstraction levels (Spring 1992). Although cognitive rehabilitation can take a number of forms, it generally involves repetitive, laboratory-based exercises, directly related to the cognitive process being trained. This might or might not involve equipment such as cognitive rehabilitation software, videotaped training material and auditory and visual stimuli. Some researchers test the efficacy of cognitive rehabilitation by comparing results from participants who have been trained to complete the required tasks against results from participants who have completed the tasks with no prior training. Others use the same method of comparing one trained group against one untrained group, but using different laboratory-based tasks for the training process than those employed in the performance assessments. A number of theorists also have argued that the generalisation of cognitive rehabilitation should be evaluated by measuring the impact of training on day-to-day activities and behaviours requiring the same processes (Spring 1992). This review sought to clarify what impact cognitive rehabilitation has on performance in the specific domain of interest and on broader but more clinically relevant outcome measures related to symptomatology and general functioning. O B J E C T I V E S The primary objective of this review was to determine whether cognitive rehabilitation is an effective treatment for people with schizophrenia and related conditions. The secondary objective was to examine whether there is a differential effect for the various types of interventions used. M E T H O D S Criteria for considering studies for this review Types of studies All relevant randomised controlled trials. 2

5 Types of participants People with a diagnosis of schizophrenia or related conditions, diagnosed by any criteria. Types of interventions 1. Cognitive rehabilitation involving repetitive laboratory-based exercises to train basic-level cognitive processes such as memory, attention, speed of processing and abstraction levels. The training might or might not involve equipment such as cognitive rehabilitation software, videotaped training material and auditory and visual stimuli. Sometimes cognitive rehabilitation is combined with monetary incentives. Cognitive rehabilitation is not cognitive-behaviour therapy (CBT) which involves training strategies to modify factors such as thoughts, beliefs, and attitudes. CBT includes strategies such as those used in self-instructional training, rational emotive therapy, illness self-management, and social problem solving. A Cochrane review has recently been published on CBT with people with schizophrenia (Jones 1998). 2. Standard care, with or without a placebo cognitive rehabilitation. Placebo cognitive rehabilitation could be some interaction that may provide similar contact between patient and carer to that encountered within the experimental procedure outlined above, but that includes no elements of cognitive rehabilitation. 3. Other interventions added to standard care that may be directed at preservation of cognitive abilities, but not cognitive rehabilitation. Studies combining cognitive rehabilitation with other types of treatment have not been included in this review at this point. Types of outcome measures 1. General level of functioning (various types of living skills, social skills, academic skills etc.): 1.1 no improvement or deterioration on the measure of level of functioning; 1.2 average change (endpoint - baseline) on the measure of level of functioning; 1.3 average endpoint score on the measure of level of functioning. 2. Mental state (hallucinations, delusions, negative symptoms etc.): 2.1 no improvement or deterioration on the measure of mental state; 2.2 average change (endpoint-baseline) on the measure of mental state; 2.3 average endpoint score on the measure of mental state. 3. Specific cognitive domain (memory, attention, speed of processing, abstraction levels etc.): 3.1 no improvement or deterioration on the specific cognitive domain; 3.2 average change (endpoint-baseline) on the specific cognitive domain; 3.3 average endpoint score on the specific cognitive domain. 4. Acceptability of the treatment: 4.1 the number of people who dropped out during the trial; 4.2 satisfaction or lack thereof with care provided; 4.3 average satisfaction score. 5. Adverse effects: 5.1 number of people who had any adverse effect. 6. Quality of life: 6.1 quality not improved or deteriorated; 6.2 average quality change; 6.3 average quality endpoint score. 7. Costs: 7.1 average direct costs; 7.2 average indirect costs. Outcomes were divided into short term (less than three months), medium term (3-12 months) and long term (more than one year). Search methods for identification of studies Electronic searches Relevant randomised trials were identified by searching several electronic data bases (Biological Abstracts, the Cochrane Schizophrenia Group s Register of Trials, EMBASE, LILACS, MEDLINE, PsycLIT and SCISEARCH). 1. BIOLOGICAL ABSTRACTS (January 1982 to 1997) was searched using the Cochrane Schizophrenia Group s phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase: [and (cognitiv* near (rehabilitati* or educat*)) or cognitive-rehab* or cognitive-educat* ] 2. COCHRANE SCHIZOPHRENIA GROUP S REGISTER (March 2000) was searched using the phrase: [(cognitiv* and (rehabilitati* or educat*)) or cognitive-rehab* or cognitive-educat* or #42=471] 3. EMBASE (January 1980 to July 1997) was searched using the Cochrane Schizophrenia Group s phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase: [and (cognitiv* near (rehabilitati* or educat*)) or cognitive-rehab* or cognitive-educat* or explode PSYCHOMOTOR PERFOR- MANCE/all] 4. MEDLINE (January 1966 to 1997) was searched using the Cochrane Schizophrenia Group s phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase: [and (cognitiv* near (rehabilitati* or educat*)) or cognitive-rehab* or cognitive-educat* or (cognitiv* and explode REHABILI- TATION/all) or explode PSYCHOMOTOR PERFORMANCE/ all] 3

6 5. PsycLIT (January 1974 to 1997) was searched using the Cochrane Schizophrenia Group s phrase for randomised controlled trials and schizophrenia (see Group search strategy) combined with the phrase: [and (cognitiv* near (rehabilitati* or educat*)) or cognitive-rehab* or cognitive-educat* or explode COGNITIVE REHABILITA- TION/all) 6. SCISEARCH - Science Citation Index Each of the included studies was sought as a citation on the SCISEARCH database. Reports of articles that have cited these studies were inspected in order to identify further trials. Searching other resources 1. Reference searching The references of all identified studies were inspected for more studies. 2. Personal contact. The first author of each included study was contacted for information regarding unpublished trials. Data collection and analysis 1. Selection of trials Two reviewers (RH, JM) independently inspected the abstract of each reference identified by the search to see if the study was likely to be relevant. When it was unclear from an abstract whether a study was a randomised trial or if there was disagreement between the two reviewers, the full article was obtained. The article was then inspected independently by the two reviewers to assess its relevance to the review. Disputes about the relevance of specific papers to the review were resolved by discussion. For studies for which there was insufficient information for reviewers to evaluate their relevance to the review, the first author of each study was contacted for further information. 2. Assessment of methodological quality The methodological quality of the trials included in this review was assessed using the criteria described in the Cochrane Handbook (Mulrow 1997) and the Jadad Scale (Jadad 1996). The former is based on the evidence of a strong relationship between the potential for bias in the results and allocation concealment, and is defined below: A. Low risk of bias (adequate concealment of treatment allocation) B. Moderate risk of bias (some doubt about the concealment of treatment allocation) C. High risk of bias (inadequate concealment of treatment allocation) The Jadad scale measures a wider range of factors that impact on the quality of the trial. It includes three items: A. Was the study described as randomised? B. Was the study described as double-blind? C. Was there a description of withdrawals or drop outs? Trials were included if they met criteria A or B of the Handbook. Also, a cut-off of two points on the Jadad scale was used to check the assessment made using the Handbook criteria. However, the Jadad scores were not used to exclude trials from this review. 3. Data collection Both reviewers independently extracted the data from the included trials. Again, any disagreement was discussed, the decisions documented and, where necessary, the first authors of the studies contacted for clarification. Reasons for excluding papers are given in the table of excluded studies. 4. Data synthesis 4.1 Incomplete data If, for a given outcome, more than 50% of participants were not accounted for, the results have not been presented, because the reviewers considered such data impossible to interpret. 4.2 Intention to treat analysis Unless a study clearly reported the reasons for leaving the study early, the reviewers assumed that participants who dropped out had no change in cognitive ability. When insufficient data were provided to identify the original group size (prior to leaving the study), the first authors were contacted. The numbers of patients who completed each study were used to test the sensitivity of the results, by comparing those trials that used intention-to-treat analysis with those that did not. 4.3 Dichotomous outcomes These were analysed by calculating odds ratios (OR) for each trial with the uncertainty in each result being expressed using 95% confidence intervals (CI). The odds ratios from the different trials were combined using the Peto methods of meta-analysis. When overall results were significant, the number needed to produce (or prevent) one outcome was calculated by combining the overall odds ratio with an estimate of the prevalence of the event in the control group of the trials. 4.4 Continuous data For continuous outcomes a weighted mean difference (WMD) between groups was estimated. If heterogeneity was found (see section 5) a random effects model was used Normal distribution - data on continuous outcomes are frequently skewed, the mean not being the centre of the distribution. The statistics for meta-analysis are thought to be able to cope with some skew, but were formulated for parametric data. To avoid this potential pitfall, the following standards were applied to all data before inclusion: (i) standard deviations and means were reported or obtained from authors; and (ii) for data with finite limits, such as endpoint scale data, the standard deviation (SD), when multiplied by two, was less than the mean. Otherwise the mean is unlikely to be an appropriate measure of the centre of the distribution (Altman 1996). The reviewers reported data which did not meet the first or second standard in the other data tables. For change data (endpoint minus baseline), the situation is even more problematic. In the absence of individual patient data, it is impossible to know if data are skewed, though likely. After consult- 4

7 ing the ALLSTAT electronic statistics mailing list, the reviewers presented change data in Metaview in order to summarise available information. In doing this, the reviewers assumed that either the data were not skewed or that the analyses could cope with the unknown degree of skew. Without individual patient data, it is impossible to test this assumption. Where both change and endpoint data were available for the same outcome category, only endpoint data were presented. The reviewers acknowledge that by doing this much of the published change data were excluded, but argue that endpoint data is more clinically relevant and that if change data were to be presented along with endpoint data it would be given undeserved, equal, prominence. Authors of studies reporting only change data are being contacted for endpoint figures Scale data - a wide range of rating scales are available to measure outcomes in mental health trials. These scales vary in quality and many are poorly validated. It is generally accepted that measuring instruments should have the properties of reliability (the extent to which a test effectively measures anything at all) and validity (the extent to which a test measures that which it is supposed to measure). Before publication of an instrument, most scientific journals insist that both reliability and validity be demonstrated to the satisfaction of referees. The reviewers decided, as a minimum standard, to exclude data from unpublished rating scales. In addition, the rating scale should have been either: (i) a self report; or (ii) completed by an independent rater or relative. More stringent standards for instruments may be set in future editions of this review. Continuous data may be presented from different scales, rating the same outcome. In this event, the reviewers presented all data without summation and inspected the general direction of effect. Where possible, reviewers entered data in such a way that the area to the left of the line of no effect indicated a favourable outcome for cognitive rehabilitation. 5. Heterogeneity Heterogeneity in the results of the trials was assessed both by inspection of graphical presentations and by calculating a Chi-square test of heterogeneity. Five possible reasons for heterogeneity were pre-specified. The reviewers proposed that responses would differ according to (i) difference in the quality of the trial; (ii) the different types of cognitive rehabilitation used; (iii) whether monetary incentives were used; and (iv) the baseline levels of symptoms and cognitive functioning of participants. These were to be assessed by looking at separate subgroups of trials. 6. Addressing publication bias Data from all selected trials were to be entered into a funnel graph (trial effect versus trial size) in an attempt to investigate the likelihood of overt publication bias. R E S U L T S Description of studies See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies. 1. Excluded Most studies of cognitive rehabilitation were excluded because there was no control group; many were single case studies. Without a plausible control condition, it was impossible to distinguish the specific benefits of cognitive rehabilitation from the non-specific treatments effects which follow most psychosocial interventions for people with schizophrenia (Spaulding 1992). Five studies were excluded because they compared two different forms of cognitive rehabilitation. These trials were either based on the assumption that cognitive rehabilitation is an effective treatment and that some forms might be more effective than others (Corrigan 1995, Young 1995); or tested different forms of reinforcement in the context of a cognitive rehabilitation intervention (Bellack 1990, Summerfelt 1991). Trials were also excluded because cognitive rehabilitation had been administered in conjunction with other forms of treatment, making it impossible to evaluate the specific contribution of cognitive rehabilitation to any improvement in participants. 2. Awaiting assessment Three studies are awaiting assessment. Funke 1989 is written in German and requires translation before it can be fully assessed for inclusion. Van der Gaag 1994 is awaiting clarification of how participants were allocated to intervention and comparison groups and whether cognitive rehabilitation was administered alone or was combined with another form of treatment. A randomised controlled trial by Spaulding 1998 awaits clarification of the specific treatments. 3. Ongoing Two authors have referred to having begun controlled trials of cognitive rehabilitation (Borg 1998, Jaeger 1992b). To date, we have no information on the status of these studies or any data generated by them. 4. Included Three studies met the inclusion criteria (Medalia 1998, Tompkins 1995, Wykes 1999). 4.1 Methods Although all three trials were randomised, none reported the method by which this was achieved. Because of the nature of the intervention, it is not practical to attempt to keep the participants blind to treatment allocation. In Medalia 1998, a rater blind to the group allocation of participants scored one of the outcome measures, the BPRS. In Wykes 1999, symptoms were assessed by a sole rater who was blind to treatment allocation, and many of the cognitive measures were computer driven and therefore scored without bias. All studies used a parallel group design. 4.2 Participants All three studies were small. The total number of people randomised to cognitive rehabilitation or control is currently 107. This renders even this meta-analysis underpowered to detect meaningful, real differences between groups. Subtle differences, 5

8 picked up by use of measures and scales, may be possible to detect, even in such small trials, but the clinical utility of such data is questionable. If cognitive rehabilitation was postulated to produce dramatic effects compared to the controls, small studies would have been adequate. This, however, was never likely. All studies involved participants whose schizophrenia had been diagnosed using operationalised criteria (either DSM-III, DSM-III- R, or DSM-IV). All participants in Medalia 1998 and Tompkins 1995 were inpatients, while those in Wykes 1999 were on an outpatient basis. Two studies specified that participants should have an identified cognitive impairment (Medalia 1998, Wykes 1999) Two studies reported the sex of participants, the majority of whom were male (Medalia 1998, Wykes 1999). In the two studies that reported participants ages, the mean ages were from years (Medalia 1998, Wykes 1999). All three studies reported that participants were on neuroleptic medication (Medalia 1998, Tompkins 1995, Wykes 1999). Psychiatric history was reported in a variety of forms: minimum hospitalisation time (Medalia six weeks), mean length of hospitalisation (Tompkins days) and time since first contact with psychiatric services (Wykes CR group - 59%>10 years, Control group - 81%>10 years). All people in Wykes 1999 had a minimum of two years contact with psychiatric services. In some studies, participants were excluded if they had a history of other neurological conditions (Medalia 1998, Tompkins 1995, Wykes 1999), major drug or alcohol abuse (Wykes 1999) or mental retardation (Medalia 1998). 4.3 Interventions Cognitive rehabilitation In Tompkins 1995, cognitive rehabilitation involved instructions in how to complete measures that were also the outcome measures: WAIS-R tasks (object assembly, picture arrangement) and the Weschler Memory Scale - Revised (logical and visual memory - immediate and delayed). Medalia 1998 used the Orientation Remedial Module (ORM), a computerised program developed for people with acquired brain injury as the experimental intervention. The ORM involves a behavioural learning format that shapes and reinforces attentive behaviour. In Wykes 1999, the cognitive rehabilitation was administered from a manual (Delahunty 1993a). Participants practised oculomotor, perceptual, fine motor and conceptual tasks of graded difficulty. The training process involved three steps: (i) overt demonstration of information processing by the therapist; (ii) overt use of the methods by the patient; and (iii) covert use ditto Control interventions Tompkins 1995 employed a comparison condition involving administering the same outcome measures as given the experimental group. Whereas cognitive rehabilitation involved giving instructional cues for completing the outcome measures, the control did not involve such training. The intervention and comparison were administered in one session. In Medalia 1998 the comparison condition involved watching National Geographic documentaries and the cognitive rehabilitation and comparison condition each involved three 20-minute sessions per week for six weeks. In Wykes 1999 the comparison condition was intensive occupational therapy in which participants were involved in activities such as relaxation, assertiveness training, life diary, comprehension of social information, and role playing. Both experimental and control interventions involved one-hour daily sessions over 40 days with participants expected to attend between three and five days per week. It is unclear from this description how much time each participant spent in treatment. 4.4 Outcomes Overall Most outcome measures used in the included studies were measures of specific cognitive domains. In two studies, the training tasks were the same as the outcome measures. Improvement on the training task would only show that participants were able to learn a very specific skill. Even if participants improve on cognitive measures that are different from the training task, researchers still need to assess whether these skills generalise to clinically-relevant tasks such as living skills and that these benefits endure over time. Also, only one continuous outcome measure was used in more than one study. Future research should use a limited number of cognitive and clinical measures with established psychometric properties to enhance the ability to meta-analyse and understand their data Scales a Not specific to cognitive functioning Brief Psychiatric Rating Scale (Overall 1962) The BPRS is a 16-item scale used to measure positive symptoms associated with severe psychiatric disorder. The symptoms are rated from one, not present to seven, extremely severe and then summed to give a total score of psychiatric symptomatology. This was the only continuous outcome measure used in more than one study (Medalia 1998, Wykes 1999). Present State Examination (Wing 1974) This is a clinician-rated scale measuring mental status. One hundred and forty symptom items are rated and combined to give various syndrome and sub-syndrome scores. Higher scores indicate greater clinical impairment. This scale was used in Wykes Rosenberg Self-Esteem Scale (Rosenberg 1973) This scale is a ten-item self-report measure. Each item involves a statement about how the respondent feels about him or herself ( I feel that I have a number of good qualities ) or aspects of his or her functioning ( I feel that I can t do anything right ). Respondents rate each item on a Likert scale from almost always true to never true. This scale was used in Wykes Social Behaviour Schedule (SBS, Wykes 1986) The SBS is a 21-item scale designed to assess a range of areas of functioning of people with long-term mental illness. The scale covers areas such as social behaviour and communication, self care, and inappropriate behaviour. The respondent s behaviour on each item during the previous month is scored by someone familiar with 6

9 him or her. Each item is rated on a five-point Likert scale, with higher scores indicating greater deficits. Wykes 1986 found the SBS to have high levels of inter-rater, inter-informant, and intersetting reliability, and some evidence of discriminative validity. The SBS was used in Wykes b Cognitive functioning No measure of a cognitive domain was used in more than one study. Continuous Performance Test (Rosvold 1956) - computerised version This is a measure of deficits of attention. The test requires people to press the space bar whenever the target letter A appears on the screen. There are 200 target letters embedded in a series of 600 letters presented randomly, one at a time for 0.83 seconds, in a 10-minute presentation. Interdisplay time is 0.17 seconds. Scores are obtained for three outcome variables: number of right letter detections, number of left letter detections, and absolute percent correct. This test was used only in Medalia Controlled oral word fluency test (Spreen 1991) This task is a measure of cognitive flexibility. Participants are required, within one minute, to provide as many different words as possible beginning with a particular letter. This test was used only in Wykes Digit Span (Wechsler 1991) This task is a measure of memory and is a subtest of the WAIS-R. It was used in Wykes Dual Span (Della Sala 1995) This task is a measure of working memory. It measures the ability of the participants to perform two tasks at the same time, i.e. visual tracking and remembering numbers presented verbally. This task was used in Wykes Hayling Sentence Completion Task (Burgess 1996) This task is a measure of cognitive flexibility. In the first part participants are asked to identify a word in order to complete a sentence correctly. In the second part of the task, they are asked to identify a word that completes the sentence incorrectly. This task was used only in Wykes Response inhibition (Wykes 1990) This task measures the ability to inhibit a previously learned response and to learn a novel one. Initially in this task, participants have to press a button adjacent to a light. Later in the task, they have to press a button that is not adjacent to the light. This task was used in Wykes Sentence Span (Daneman 1980) This task is a measure of memory. Participants are required to remember the last word of sentences in groups of sentences that have been read to them. This test was used in Wykes Six Elements (Burgess 1991) This task is a measure of planning ability. Participants are given 10 minutes to complete three two-part tasks (sections A & B). However, they cannot perform section B immediately following section A of the same task. This task was used in Wykes Stroop Neuropsychological Screening Test (REF) This test is a measure of cognitive flexibility. Participants are asked to name the ink colour of printed words that describe a different colour. This test was used in Wykes Tower of London (Morris 1995) This task is a measure of planning ability. It is a computerised task in which participants move discs on a touch screen in an attempt to make as few moves as possible in a planning task. This task was used in Wykes Trails (Reitan 1958) Trails is another measure of cognitive flexibility. It is a computerised version of Trails A and B of the Halstead-Reitan battery. It was used only in Wykes Wechsler Adult Intelligence Scale - Revised (WAIS-R; Wechsler 1991) The WAIS-R is a measure of intelligence. It is a test battery comprising six verbal tests (information, comprehension, arithmetic, similarities, digit span, vocabulary) and five performance tests (digit symbol, picture completion, block design, picture arrangement, object assembly). Each test within the WAIS-R can be used on its own, and has established reliability and validity (Lezak 1995). WAIS-R task - Object Assembly (OA, Peck 1987) The OA subtest of the WAIS-R is a measure of visual organisation and construction ability. Participants are required to unscramble four pictures of cut-up objects within a specified time. It was used only in Tompkins WAIS-R task - Picture Arrangement (PA, Peck 1987) The PA subtest of the WAIS-R is a 10-item meaure of the ability to detect nonverbal social cues and to think in a logical and sequential manner. The items involve four to six cards which depict part of a story; participants are required to arrange the cards in order within specified times. This was used only in Tompkins Wechsler Memory Scale-Revised sub-tests (Wechsler 1987) Wechsler Memory Scale-Revised - logical memory (LM) sub-set (immediate and delayed) The LM subtest of the Wechsler Memory Scale-Revised measures both immediate and delayed verbal memory. The examiner orally presents two brief stories to the person who is required to repeat the stories immediately following their presentation and again after 30 minutes. High scores indicate better functioning. This was used only in Tompkins Wechsler Memory Scale-Revised - visual memory (VM) sub-set (immediate and delayed) The VM subtest of the Wechsler Memory Scale-Revised measures both immediate and delayed memory of visual information. The participant is shown three designs, one at a time, for 10 seconds each. After the presentation of each design, he or she is required to immediately draw the design from memory, and then again 30 minutes later. High scores indicate better functioning. This was used only in Tompkins Wisconsin Card Sort Test (Milner 1963) 7

10 This task is a measure of cognitive flexibility, abstract reasoning, and the ability to monitor one s behaviour and learn from feedback. It is specifically sensitive to lesions of the frontal lobe of the brain. Participants have to categorise simple geometric forms by the shape, colour, or number of the objects presented. A trial card is presented to the participant who must select a target card that matches the trial card according to one critical feature (shape, colour, or number). After the participant has successfully matched 10 consecutive trials (a category), the critical feature is changed without warning. Final scores are given for the number of categories completed and the number of preservative errors made. This test was used in Wykes Visual Span (Lezak 1995) Visual span is a validated robust measure which tests the immediate recall of visually presented sequences. Participants are required to reproduce from memory increasingly complex figures presented on a grid (Wykes 1999). Risk of bias in included studies 1. Randomisation Only one study elaborated on the method of randomisation used. Medalia 1998 stated that individuals with like [test] rankings were paired into blocks of two and each was then randomly assigned. In the other two included studies randomisation was stated, but not described. 2. Blinding Medalia 1998 stated that one of the two measures used in their study, the BPRS, was scored by a rater blind to group status. In Wykes 1999, symptoms were assessed by a sole rater who was blind to treatment allocation and many of the cognitive measures were computer driven, and therefore scored without bias. 3. Follow-up Although the specified intention had been to examine outcomes in the short term (less than three months), medium term (3-12 months), and long term (longer than one year), the only available results were short-term. No study conducted a follow-up assessment beyond the period of intervention. Two studies mentioned attrition. Six of the 60 people in Medalia 1998 dropped out of the study because of psychotic decompensation. Although the trialists did not specify which group they had been in, three participants appear to have dropped out of each group. In Wykes 1999, three participants dropped out of cognitive rehabilitation and one out of the comparison group. There was no mention of why these people left early. The methodology of Tompkins 1995 strongly limited the likelihood of attrition; the outcome measure was administered once and this was in conjunction with the single training session. Effects of interventions 1. The search The searching identified about 500 citations that were inspected. Forty-three papers were acquired and read. Three studies were finally included. 2. COGNITIVE REHABILITATION versus PLACEBO OR NO INTERVENTION 2.1 Mental state The Brief Psychiatric Rating Scale (BPRS) was the only continuous outcome measure used (Medalia 1998). The process of cognitive rehabilitation did not have any immediate effect on mental state when compared to a placebo intervention. 2.2 Leaving the study early Six people in Medalia 1998 dropped out of the study because of psychotic decompensation (total n=60). There was no mention, however, of which group they had been allocated to. Based on published data, the reviewers assumed that three people from each arm (30 in each arm) had dropped out. Cognitive rehabilitation is not obviously an intervention that people with seriously mental illness find more difficult to tolerate than the placebo approaches (OR , n=84, N=2). 2.3 Specific cognitive domains Attention Medalia 1998 reported right and wrong letter detection and total percent correct as measures of attention. No data suggests an immediate measurable effect of the cognitive rehabilitation Memory Logical and visual memory were both tested in Tompkins 1995 and cognitive rehabilitation has no clear effects in the very short term WAIS subset tests Data from object assembly and picture arrangement tests are reported but no difference between cognitive rehabilitation and the placebo approach is demonstrated. It remains unclear if this is a function of underpowering of studies and meta-analysis, rather than a true equivocal effect on which confidence can be placed. 3. COGNITIVE REHABILITATION versus OCCUPA- TIONAL THERAPY 3.1 Mental state Wykes 1999 measured mental state in two ways - using the Brief Psychiatric Rating Scale and the Present State Examination. The data presented in reports of this study suggest no differences in the effects of the two interventions in the very short term. 3.2 Social behaviour There are no discernible, measurable effects of either intervention on social behaviour. 3.3 Leaving the study early Three out of 17 participants dropped out of the cognitive rehabilitation group and one out of 16 in the comparison group (OR 3.21 CI ). 3.4 Self esteem Wykes 1999 measured this variable using the Rosenberg Self Esteem Scale. Those allocated to the cognitive rehabilitation group 8

11 have statistically significantly greater self esteem than people in the occupational therapy group, at the end of testing (MD 6.3 CI ). 3.5 Specific cognitive domains Cognitive flexibility For the tests of cognitive flexibility (controlled oral fluency test, Stroop test, Hayling sentence completion test, response inhibition, trials and the Wisconsin card sorting test) there are no clear differences between those allocated to cognitive rehabilitation and people in the occupational therapy group. Wide confidence intervals were obtained and the equivalence of the two interventions cannot be assumed Memory Sentence span, visual span and dual span are not clearly different between the groups Planning Planning was measured using the Six Elements test and the Tower of London test. Again, there are no differences demonstrated between groups WAIS subtest - digit span Wykes 1999 use the digit span sub-test of the WAIS. There is no difference between the groups. As with all the equivocal results reported above, wide confidence intervals were obtained and the equivalence of the two interventions cannot be assumed. 4. Other analyses Although the intention had been to examine outcomes in the short term (less than three months), medium term (3-12 months), and long term (longer than one year), the only available results were immediate term. A lack of adequate data also precluded the use of funnel analyses to estimate potential publication biases. Based on published data, the reviewers were only able to extract a limited amount of data from three studies. This is not entirely indicative of the small numbers of studies that have evaluated cognitive rehabilitation. It is indicative, however, of some consistent methodological issues with those studies. As indicated in the table of excluded studies, the majority of studies of cognitive rehabilitation were excluded because there was no relevant control group; many were single case studies. Without a plausible control condition, it is impossible to distinguish the specific benefits of cognitive rehabilitation from the non-specific treatments effects that appear to follow most psychosocial interventions with people with schizophrenia (Spaulding 1992). Five studies were excluded because they compared two forms of cognitive rehabilitation and had no control group relevant to this review. These studies were either based on the assumption that cognitive rehabilitation is an effective treatment and some forms might be more effective than others (Corrigan 1995, Young 1995), or they were testing different forms of reinforcement in the context of a cognitive rehabilitation intervention (Bellack 1990, Summerfelt 1991). In the former, given that cognitive rehabilitation does not have established efficacy, there seems little point in studying the efficacy of one form over another. Studies were also excluded because cognitive rehabilitation had been administered in conjunction with other forms of treatment. For example, Brenner s ( ) oft-cited Integrated Psychological Therapy (IPT) involves cognitive rehabilitation, cognitive behavioural therapy and skills training. In such cases, it is impossible to evaluate the specific contribution of cognitive rehabilitation. 1.3 Small studies D I S C U S S I O N 1. General 1.1 Definitions The definition of cognitive rehabilitation used was relatively broad. The reviewers categorised instruction on the rules required to complete particular tests of cognitive skills as a form of cognitive rehabilitation (Tompkins 1995). Some might suggest that this is in fact cognitive behavioural therapy and not cognitive rehabilitation because participants are instructed in a strategy to deal with a situation. The reviewers regarded such interventions as cognitive rehabilitation because, rather than intending to modify factors such as thoughts, beliefs and attitudes (as with their definition of cognitive behavioural therapy), these rules were used to train basic level cognitive processes (in accordance with their definition of cognitive rehabilitation). 1.2 Few studies The randomised trial-derived evidence for cognitive rehabilitation for schizophrenia currently consists of a handful of studies. The three included studies had small numbers of participants (ranging from 24-55) and no included trial mentioned how sample size was calculated. Trials should have adequate power to detect clinically moderate effects. A recent review has warned against placing too much emphasis on data derived from a small number of small trials (Egger 1995). 1.4 Short studies The included studies involved a broad range of duration and intensity of cognitive rehabilitation interventions. These ranged from one treatment session in total by Tompkins 1995 to three 20- minute sessions a week for six weeks by Medalia For an intervention to have had an adequate trial, participants need to have had an adequate dose of that intervention. Even though it is difficult to say at this stage what an adequate dose might be, it is highly questionable whether a population with memory deficits, like people with schizophrenia, would benefit from a very brief intervention. The duration of follow-up in all trials was very short. 9

12 There are no data on the effects, or lack of effects of this intervention for periods beyond six weeks and none after the intervention had stopped. 1.5 Studies of unclear clinical relevance The great majority of outcomes used in the included studies were measures of specific cognitive domains. In two studies, the training tasks were the same as the outcome measures. Improvement on the training task would only show that participants were able to learn a very specific skill. It would not show that the underlying cognitive processes required for the tasks have improved, or whether the skills being measured generalise to other tasks which require the same basic-level cognitive processes. An important factor which limited the analysis was that only one continuous outcome measure was used in more than one study, or had similar psychometric properties to any other measure. Future research should use a limited number of cognitive and clinical measures with established psychometric properties to enhance the ability to analyse data. Similarly, even if participants are able to improve on measures that are different from the training task, researchers still need to assess whether these skills generalise to more clinically relevant tasks, such as living skills and symptomatology. Ultimately, for cognitive rehabilitation to be regarded as having clinical value, it must be shown to have a positive and lasting impact which extends beyond measures of specific cognitive domains. 1.6 Poorly reported studies The power of this review was limited by incomplete reporting of methodology and data. Specific details about methods of group allocation, concealment of treatment and attrition are important but were inadequately reported. Similarly, measures, particularly those without established psychometric properties, should be well described. Averaging results across groups without presenting the standard deviation prevents data from being used for systematic reviews. The quality of studies appears, however, to be improving. Wykes 1999 has dealt with the methodological issues identified above most comprehensively. For example, this study specified treatment concealment, used measures with established psychometric properties, and administered cognitive rehabilitation independently of other treatment. Wykes 1999 also measured outcomes in all domains: general functioning, symptomatology, specific cognitive domains, and acceptability of treatment. 2. COGNITIVE REHABILITATION versus PLACEBO OR NO INTERVENTION Analyses revealed data within this comparison to be inconclusive - providing no evidence for or against cognitive rehabilitation as a treatment for schizophrenia. Measures of mental state and specific cognitive domains demonstrated no effects. Cognitive rehabilitation was as acceptable experience as the placebos offered. 3. COGNITIVE REHABILITATION versus OCCUPA- TIONAL THERAPY Again, analyses revealed data within this comparison to be inconclusive - providing little evidence for or against cognitive rehabilitation when compared to occupational therapy as a treatment for schizophrenia. Measures of mental state and specific cognitive domains demonstrated no effects. Cognitive rehabilitation was as acceptable an experience as occupational therapy. Self esteem, however, was greater in those allocated to the cognitive rehabilitation group. The reviewers are unclear as to the clinical significance of a mean six-point difference on the Rosenberg Self Esteem Scale and how this may translate into mental state, functioning and behaviour. This could also be a function of random error and should be re-investigated in any subsequent studies. A U T H O R S C O N C L U S I O N S Implications for practice This review was unable to locate sufficient trial-derived evidence to inform clinicians, managers and recipients of care about the role of cognitive rehabilitation in practice. If clinicians wish to use cognitive rehabilitation, the experimental nature of the intervention should be clearly explained to the patient. Implications for research Further studies It is likely that additional studies are being undertaken on this experimental intervention. Given that there is insufficient quality research to say whether cognitive rehabilitation is effective or not, there is a need for further research to establish whether or not it can remediate the cognitive deficits common in this population. This is important because of the impact that these deficits have on the functioning of people with schizophrenia, and the implications of such findings regarding the neural plasticity of people with schizophrenia. If additional studies are being planned, trials should: (i) use designs that allow the specific effects of cognitive rehabilitation to be demonstrated (Brenner 1994, for example, combined cognitive rehabilitation with other treatments with proven benefit [psychoeducation] which prevented any benefits accruing from cognitive rehabilitation alone from being assessed); (ii) have adequate power to detect clinically important effects; (iii) be of sufficiently long duration to reassure readers that a lasting benefit is occurring; (iv) consider what is an adequate dose of cognitive rehabilitation for this population (the duration of intervention ranged from one session in total to three 20-minute sessions per week for six weeks in the included studies); (v) use measures that are clinically relevant and a limited number of clinical and cognitive measures with established psychometric properties; and (vi) clearly and completely report or make data available. 10

13 A C K N O W L E D G E M E N T S We thank Clive Adams for his invaluable editorial support. R E F E R E N C E S References to studies included in this review Medalia 1998 {published data only} Medalia A, Aluma M, Tryon W, Merriam AE. Effectiveness of attention training in schizophrenia. Schizophrenia Bulletin 1998;24: Tompkins 1995 {published data only} Tompkins LM, Goldman RS, Axelrod BN. Modifiability of neuropsychological dysfunction in schizophrenia. Biological Psychiatry 1995;38: Wykes 1999 {published data only} Wykes T, Reeder C, Corner J, Williams C, Everitt B. The effects of neurocognitive remediation on executive processing in patients with schizophrenia. Schizophrenia Bulletin 1999;25: References to studies excluded from this review Adams 1981 {published data only} Adams HE, Malatesta V, Brantley PJ, Turkat ID. Modification of cognitive processes: a case study of schizophrenia. Journal of Consulting and Clinical Psychology 1981;49: Ahmed 1994 {published data only} Ahmed M, Goldman JA. Cognitive rehabilitation of adults with severe and persistent mental illness: a group model. Community Mental Health Journal 1994;30: Bellack 1990 {published data only} Bellack AS, Mueser KT, Morrison RL, Tierney A, Podell K. Remediation of cognitive deficits in schizophrenia. American Journal of Psychiatry 1990;147: Benedict 1989 {published data only} Benedict RHB, Harris AE. Remediation of attention deficits in chronic schizophrenic patients: a preliminary study. British Journal of Clinical Psychology 1989;28: Benedict 1994 {published data only} Benedict RHB, Harris AE, Markow T, McCormick JA, Nuechterlein KH, Asarnow RF. Effects of attention training on information processing in schizophrenia. Schizophrenia Bulletin 1994;20: Brenner 1994 {published data only} Brenner HD. The treatment of basic psychological dysfunctions from a systemic point of view. British Journal of Psychiatry 1989;155(Supplement 5): Brenner HD, Boker W, Hodel B, Wyss H. Cognitive treatment of basic pervasive dysfunctions in schizophrenia. In: Schulz SC, Tamminga CA editor(s). Schizophrenia: scientific progress. New York: Oxford University Press, Brenner HD, Hodel B, Genner R, Roder V, Corrigan P. Biological and cognitive vulnerability factors in schizophrenia: implications for treatment. British Journal of Psychiatry 1992;161(Supplement 18): Brenner HD, Hodel B, Roder V, Corrigan P. Treatment of cognitive dysfunctions and behavioral deficits in schizophrenia. Schizophrenia Bulletin 1992;18:21 6. Brenner HD, Stramke WG, Mewes F, Liese F, Seeger G. Experiences with a specific therapy program for training cognitive and communicative skills in the rehabilitation of chronic schizophrenic patients [Erfahrungen mit einem spezifischen therapieprogramm zum training kognitiver and kommunikativer fähigkeiten in der rehabilitation chronisch schizophrener patienten]. Der Nervenarzt 1980;51: Brown 1993 {published data only} Brown C, Harwood K, Hays C, Heckman J, Short JE. Effectiveness of cognitive rehabilitation for improving attention in patients with schizophrenia. Occupational Therapy Journal of Research 1993;13: Corrigan 1995 {published data only} Corrigan PW, Hirschbeck JN, Wolfe M. Memory and vigilance training to improve social perception in schizophrenia. Schizophrenia Research 1995;17: Delahunty 1993 {published data only} Delahunty A, Morice R, Frost B. Specific cognitive flexibility rehabilitation in schizophrenia. Psychological Medicine 1993;23: Fine 1994 {published data only} Fine SB. Reframing rehabilitation: putting skill acquisition and the mental health system into proper perspective. In: Spaulding WD editor(s). Cognitive technology in psychiatric rehabilitation. Lincoln, NE: University of Nabraska Press, Finnell 1997 {published data only} Finnell A, Card J, Menditto A. A comparison of appropriate behavior scores of residents with chronic schizophrenia participating in therapeutic recreation services and vocational rehabilitation services. Therapeutic Recreation Journal 1997;31(First Quarter): Garety 1994 {published data only} Garety PA, Kuipers L, Fowler D, Chamberlain F, Dunn G. Cognitive behavioural therapy for drug-resistant psychosis. British Journal of Medical Psychology 1994;67: Goldberg 1994 {published data only} Goldberg J. Cognitive retraining in a community psychiatric rehabilitation program. In: Spaulding WD editor(s). Cognitive technology in psychiatric rehabilitation. Lincoln, NE: University of Nabraska Press,

14 Granholm 1992 {published data only} Granholm E. Processing resource limitations in schizophrenia: implications for predicting medication response and planning attentional training. In: Margolin DI editor(s). Cognitive neuropsychology in clinical practice. New York: Oxford University Press, Jaeger 1992a {unpublished data only} Jaeger J, Douglas E. Neuropsychiatric rehabilitation for persistent mental illness. Psychiatric Quarterly 1992;63: Kern 1994 {published data only} Kern RS, Green MF. Cognitive prerequisites of skill acquisition in schizophrenia: bridging micro- and macrolevels of processing. In: Spaulding WD editor(s). Cognitive technology in psychiatric rehabilitation. Lincoln, NE: University of Nabraska Press, Konen 1991 {published data only} Konen A, Neis L, Hodel B, Brenner HD. Cognitivebehavioral therapy of schizophrenia. The Integrative psychological therapies (IPT) [A propos des thérapies cognitive comportementales de la schizophrénie. Le programme intégratif de thérapies psychologiques (IPT)]. L Encéphale 1993;XIX: Michel 1998 {published data only} Michel L, Danion J-M, Grange D, Sander G. Cognitive skill learning and schizophrenia: implications for cognitive rehabilitation. Neuropsychology 1998;12: Morice 1996 {published data only} Morice R, Delahunty A. Treatment strategies for the remediation of neurocognitive dysfunction in schizophrenia. In: Pantelis C, Nelson HE, Barnes TRE editor(s). Schizophrenia: a neurological perspective. New York: John Wiley & Sons, Nisbet 1996 {published data only} Nisbet H, Siegert R, Hunt M, Fairley N. Improving schizophrenic in-patients Wisconsin card-sorting performance. British Journal of Clinical Psychology 1996;35: Perris 1992 {published data only} Perris C. A cognitive-behavioral treatment program for patients with a schizophrenic disorder. New Directions for Mental Health Services 1992;53: Reed 1992 {published data only} Reed D, Sullivan, ME, Penn DL, Stuve P, Spaulding WD. Assessment and treatment of cognitive impairments. New Directions for Mental Health Services 1992;53:7 19. Spaulding 1986 {published data only} Spaulding WD, Storms L, Goodrich V, Sullivan M. Applications of experimental psychopathology in psychiatric rehabilitation. Schizophrenia Bulletin 1986;12: Spaulding 1993a {published data only} Spaulding WD. Spontaneous and induced changes in rehabilitation of chronic schizophrenia. In: Cromwell RL, Snyder CR editor(s). Schizophrenia: origins, processes, treatment, and outcome. New York: Oxford University Press, Spaulding 1993b {published data only} Spaulding WD. Spontaneous and induced changes in rehabilitation of chronic schizophrenia. In: Cromwell RL, Snyder CR editor(s). Schizophrenia: origins, processes, treatment, and outcome. New York: Oxford University Press, Spaulding 1994 {published data only} Spaulding WD, Sullivan M, Weiler M, Reed D, Ricardson C, Storzbach D. Changing cognitive functioning in rehabilitation of schizophrenia. Acta Psychiatrica Scandinavica 1994;90(Supplement 384): Summerfelt 1991 {published data only} Summerfelt AT, Alphs LD. Reduction of perseverative errors in patients with schizophrenia using monetary feedback. Journal of Abnormal Psychology 1991;100: Tryssenaar 1994 {published data only} Tryssenaar J, Goldberg J. Improving attention in a person with schizophrenia. Canadian Journal of Occupational Therapy 1994;61: Trzepacz 1991 {published data only} Trzepacz P, Starratt C. A neuropsychiatric model of treatment. In: Tamminga CA, Schultz SC editor(s). Advances in neuropsychiatry and psychopharmacology. Vol. 1, New York: Raven Press, Velligan 1996 {published data only} Velligan DI, Mahurin RK, True JE, Lefton RS, Flores CV. Preliminary evaluation of cognitive adaptation training to compensate for cognitive deficits in schizophrenia. Psychiatric Services 1996;47: Vollema 1995 {published data only} Vollema MG, Geurtsen GJ, Augustinus JP, van Voorst JP. Durable improvements in Wisconsin Card Sorting Test performance in schizophrenic patients. Schizophrenia Research 1995;16: Wexler 1997 {published data only} Wexler BE, Hawkins KA, Rounsaville B, Anderson M, Sernyak MJ, Green MF. Normal neurocognitive performance after extended practice in patients with schizophrenia. Schizophrenia Research 1997;26: Young 1995 {published data only} Young DA, Freyslinger MG. Scaffolded instruction and the remediation of Wisconsin Card Sorting Test deficits in chronic schizophrenia. Schizophrenia Research 1995;16: References to studies awaiting assessment Funke 1989 {published data only} Funke B, Reinecker H, Commichau A. Frontiers of cognitive training methods for long-term schizophrenic patients [Grenzen kognitiver trainingsmethoden bei schizophrenen langzeitpatienten]. Der Nervenartz 1989;60:

15 Spaulding 1998 {published data only} Spaulding W, Reed D, Storzbach D, Sullivan M, Weiler M, Richardson C. The effects of a remediational approach to cognitive therapy for schizophrenia. In: Wykes T, Tarrier N, Lewis S editor(s). Outcome and innovation in psychological treatment for schizophrenia. New York: John Wiley & Sons, Van der Gaag 1994 {published data only} Van der Gaag M, Woonings FMJ, van den Bosch RJ, Appelo MT, Slooff CJ, Louwerens JW. Cognitive training of schizophrenic patients: a behavioural approach based on experimental psychopathology. In: Spaulding WD editor(s). Cognitive technology in psychiatric rehabilitation. Lincoln, NE: University of Nabraska Press, References to ongoing studies Borg 1998 {published data only} Rund BR, Borg NC. A cognitive training programme for schizophrenic patients. Schizophrenia Research 1998;29: 164. Jaeger 1992b {published data only} Jaeger J, Douglas E. Neuropsychiatric rehabilitation for persistent mental illness. Psychiatric Quarterly 1992;63: Additional references Altman 1996 Altman DG, Bland JM. Detecting skewness from summary information. BMJ 1996;313:1200. [: COGR020600] Bellack 1992 Bellack AS. Cognitive rehabilitation for schizophrenia: is it possible? Is it necessary?. Schizophrenia Bulletin 1992;18: Burgess 1996 Burgess PW, Shallice T. Response suppression, initiation and strategy use following frontal lobe lesions. Neuropsychologia 1996;34(4): [: ] Daneman 1980 Daneman M, Carpenter PA. Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior 1980;19(4): Delahunty 1993a Delahunty A, Morice R. The training programme for the remediation of cognitive deficits in schizophrenia. Albury, NSW: New South Wales Department of Health, Della Sala 1995 Della Sala S, Baddeley A, Papagno C, Spinnler H. Dualtask paradigm: a means to examine the central executive. In: Grafman J, Holyoak KJ, Boller F editor(s). Structure and functions of the human prefrontal cortex: annals of the New York Academy of Sciences. Vol. 769, New York: New York Academy of Sciences, 1995: [: (hardcover); (paperback)] Egger 1995 Egger M, Davey Smith G. Misleading meta-analysis. BMJ 1995;310: Gold 1993 Gold JM, Harvey PD. Cognitive deficits in schizophrenia. Psychiatric Clinics of North America 1993;16(2): [: ] Jadad 1996 Jadad AR, Moore A, Carroll D, Jenkinson C, Reynolds JM, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinicalt trials: is blinding necessary?. Controlled Clinical Trials 1996;17:1 12. Jones 1998 Jones C, Cormac I, Mota J, Campbell C. Cognitive behaviour therapy for schizophrenia. Cochrane Database of Systematic Reviews 1998, Issue 4. [DOI: / CD pub2] Lezak 1995 Lezak MD. Neurological assessment. 3rd Edition. New York: Oxford University Press, Morris 1995 Morris RG, Rushe T, Woodruffe PWR, Murray RM. Problem solving in schizophrenia: a specific deficit in planning ability. Schizophrenia Research 1995;14(3): Mulrow 1997 Mulrow CD, Oxman AD. Cochrane Collaboration Handbook [updated 1 September 1997] The Cochrane Collaboration, Cochrane Database of Systematic Reviews. Oxford: Update Software; Updated quarterly, Overall 1962 Overall JE, Gorham DR. The Brief Psychiatric Rating Scale. Psychological Reports 1962;10: [: ] Peck 1987 Peck EA, Stephens V, Martelli MF. A descriptive summary of essential neurological tests. In: Hartlage LC, Asken MJ, Hornsby JL editor(s). Essentials of neurological assessment. New York: Springer, 1987: Reitan 1958 Reitan RM. Validity of the Trail Making Test as an indicator of organic brain damage. Perceptual and Motor Skills 1958; 8: Rosenberg 1973 Rosenberg M. Self-esteem scale. In: Robinson JP, Shaver PR editor(s). Measures of social psychological attitudes. Revised. Ann Arbor, MI: Institute for Social Research, University of Michigan, Rosvold 1956 Rosvold HE, Mirsky AF, Sarason I, Bransome ED Jr, Beck LH. A continuous performance test of brain damage. Journal of Consulting Psychology 1956;20:

16 Spaulding 1992 Spaulding WD. Design prerequisites for research on cognitive therapy for schizophrenia. Schizophrenia Bulletin 1992;18: Spreen 1991 Spreen O, Strauss E. A compendium of neuropsychological tests: administration, norms, and commentary. New York: Oxford University Press, [: ] Spring 1992 Spring BJ, Ravdin L. Cognitive remediation in schizophrenia: should we attempt it?. Schizophrenia Bulletin 1992;18: Wing 1974 Wing JK, Cooper J, Sartorious N. The measurement and classification of symptoms. Cambridge: Cambridge University Press, Wykes 1986 Wykes T, Sturt E. The measurement of social behaviour in psychiatric patients: an assessment of the reliability and validity of the SBS Schedule. British Journal of Psychiatry 1986;148:1 11. References to other published versions of this review Hayes 2000 Hayes R, McGrath JJ. Cognitive rehabilitation for people with schizophrenia and related conditions: a systematic review and meta-analysis. Schizophrenia Research. 2000; Vol. 41, issue 1, Special Issue: Indicates the major publication for the study 14

17 C H A R A C T E R I S T I C S O F S T U D I E S Characteristics of included studies [ordered by study ID] Medalia 1998 Methods Participants Interventions Outcomes Allocation: pairs matched by test rankings and randomised - no further description. Blindness: BPRS scored by blind rater. Design: parallel. Setting: in hospital. Diagnosis: schizophrenia (DSM-III-R). History: IQ>70, impaired attention <99% correct on CPT, in hospital > 6 weeks before study, on neuroleptics, no diagnosed brain disease. N = 60.* Sex: 47 M, 13 F. Age: mean ~ 33 years (SD ~6.5). 1. Cognitive rehabilitation (ORM computer program developed for people with head injuries emphasising practice in a behavioral learning format that shapes and reinforces attentive behavior through engaging in computerized exercises. N=30.* 2. Control: viewing National Geographic documentaries with a clinician. N=30.* Both 3 X 20 minute sessions per week for 6 weeks. Mental state: BPRS. Leaving the study early. Specific cognitive domains: right letter detections, wrong letter detections, absolute percentage correct (CPT). Reaction time. Notes Jadad score = 1. * 6 participants dropped out, trialists analysed 54 (27 each group) - reviewers assumed 3 lost / group - unclear when attrition occured. Letter sent to first author for further information. Risk of bias Item Authors judgement Description Allocation concealment? Unclear B - Unclear 15

18 Tompkins 1995 Methods Participants Interventions Outcomes Allocation: randomly assigned to an evaluator who was blind to the diagnosis of the patient - no further information. Blindness: as described above - no further information. Design: parallel. Setting: in hospital. Diagnosis: schizophrenia (DSM-III-R). History: mean age of onset ~ 21 years, mean days hospitalisation pre-study ~ 240 (SD 213), no dementia, mean chlorpromazine equivalent ~ 1 gm/day. N = 24. Sex: not specified. Age: mean ~ 37 years (SD ~ 6). 1. Cognitive rehabilitation: cued condition with instrumental cues on measures of visual and semantic memory, executive function, and constructional ability. N=12.* 2. Control: performed the same neuropsychological measures without cues. N=12.* Both administered in one session. Specific cognitive domains: memory, logical and visual (Wechsler Memory Scale). Specific cognitive domains: WAIS-R tasks - object assembly, picture arrangement Notes Jadad score = 1. * Both groups trained on outcome measures - only difference being inclusion of cueing in group 1. Letter sent to first author for further information. Risk of bias Item Authors judgement Description Allocation concealment? Unclear B - Unclear Wykes 1999 Methods Participants Interventions Allocation: random - no further details. Blindness: raters blind to group assignment. Design: parallel. Setting: day treatment centre. Diagnosis: schizophrenia (DSM-IV). History: evidence of cognitive difficulties, no evidence of organic brain disease, no plans to change medication during treatment. N=23. Sex: 15 M, 8 F. Age: mean ~ 38 years. 1. Neurocognitive remediation (CR) - CR as set out in Delahunty and Morice s (1993) manual. In each session, a variety of tasks were presented to practice each of the component processes in complex planning or problem solving 16

19 Wykes 1999 (Continued) 2. Intensive Occupational Therapy (IOT) - including relaxation, assertiveness training, life diary, comprehension of social information, and role playing. Both 1-hour daily sessions over 40 days for 3 to 5 days per week Outcomes Mental state: BPRS, PSE. General functioning: SBS. Self esteem: Rosenberg Self Esteem Scale. Leaving the study early. Specific cognitive domains: Cognitive Flexibility (Hayling Sentence Completion Task, Trails, Response Inhibition, Controlled Oral Word Fluency Test, Stroop, WCST), Planning (Tower of London, Six Elements), Memory & Working Memory (Visual, Sentence,Digit, & Dual Span) Notes Jadad score = 1. Used intention to treat analysis. Risk of bias Item Authors judgement Description Allocation concealment? Yes A - Adequate Mental state scales BPRS - Brief Psychiatric Rating Scale SBS - Social Behaviour Schedule PSE - Present State Examination Cognitive measure CPT - Continuous Performance Test IQ - Intelligence quotent WCST - Wisconsin Card Sort Test Diagnostic aid DSM-III-R - Diagnostic Statistical Manual - version 3, revised General F - female M - male Characteristics of excluded studies [ordered by study ID] Study Adams 1981 Ahmed 1994 Reason for exclusion Allocation: case study, not randomised. Allocation: case series, not randomised. 17

20 (Continued) Bellack 1990 Benedict 1989 Benedict 1994 Brenner 1994 Brown 1993 Corrigan 1995 Delahunty 1993 Fine 1994 Finnell 1997 Allocation: allocated sequentially, not randomly. Participants: people with schizophrenia. Interventions: cognitive rehabilitation versus contingent reinforcement and noncontingent reinforcement, no placebo Allocation: randomly assigned - no further details. Participants: 20 people with schizophrenia (DSM-III), mean age ~30 years (SD 5.6) taking mean of 709mg chlorpromazine per day. Interventions: cognitive rehabilitation (computerised attention-training tasks - progression through task hierarchy dependent on improved performance. 11 tasks - speed of information processing & vigilance ; 14 tasks - skills in memory, concept formation & problem solving ) versus attention placebo (same tasks & attention as experimental group, but no progression criteria, equal time spent on each task) versus no treatment control. Duration of cognitive rehabilitation & attention placebo: 25 X 30 minute sessions. Outcomes: reaction time (+/- auditory distraction), specific reaction time tasks (total = 120 trials) - no usable data. Letter sent to first author for further information. Allocation: randomly assigned in sequence, quasi-randomised. Participants: people with schizophrenia. Interventions: cognitive rehabilitation versus no-treatment control Allocation: not randomised, describes Integrated Psychological Therapy Allocation: randomised - no further description. Participants: 29 people with chronic schizophrenia (DSM-IIIR), mean age ~ 50 years, mean length of stay 7 years. Interventions: Cognitive rehabilitation (Attention Process Training - a hierarchical, multilevel treatment program designed to remediate attention deficits in brain-injured persons... primarily consisting of paper-and-pencil and auditory stimuli/motor response tasks versus control group ( traditional one-to-one task-oriented occupational therapy program aimed at improving cognitive skills through task completion ). Duration of both interventions: 3 X 60 minute sessions per week for 12 weeks Outcomes: Digit span subscale of WAIS, visual span subscale of the revised memory scale, digit symbol subscale of WAIS, trail making subtests A & B of the Halstead Reitan Neurological Battery, Bay Area Functional Performance Evaluation (BaFPE). Unusual treatment of data from two groups - [because] neither treatment modality was more effective than the other.... statistical analysis was done on the combined score of the two treatment groups. Letter sent to first author for further information. Allocation: randomised. Participants: people with schizophrenia and schizoaffective disorder. Interventions: two forms of cognitive rehabilitation, no control group Allocation: case studies, not randomised. Allocation: case study, not randomised. Allocation: not randomised. 18